---
format: "aidr-story-markdown/v1"
id: "c5e280287d35fde94e30e54091177568ba665b19c2c4c79e29044b94d4973cc3"
canonical_url: "https://aidr.today/c5e28028?lang=en"
title: "MiniMax Sets Oct 23 Weight Release for M3.1 Flash AI Model"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-19T12:26:12.000Z"
category: "Open Source"
topics: ["llm","open-source","devtools"]
source_urls: ["https://huggingnews.com/ai/minimax-sets-oct-23-weight-release-for-m31-flash-ai-model-a3ac7182","https://x.com/MiniMax_AI/status/2112150867868541221","https://x.com/OpenRouter/status/2112152045461680591","https://x.com/OpenRouter/status/2112152058912755726","https://x.com/OpenRouter/status/2112153289907470804","https://x.com/kimmonismus/status/2112151168579227892","https://x.com/teortaxesTex/status/2112152366200021378","https://marketbrief.now/ai/minimax-sets-oct-23-weight-release-for-m31-flash-ai-model-a3ac7182"]
summary: "MiniMax will release the weights for its M3.1 Flash AI model, formerly known as 'Space Bunny', on Oct 23. The model runs at 3x the speed of the M3 model at one third the cost, with flat rate API pricing of $0.15 per 1 million input tokens and $0.60 per 1 million output tokens. Performance benchmarks include a 74% score on DeepSWE for coding and 91% on BabyVision for visual reasoning. During its stealth phase, MiniMax used performance feedback to cut error rates to 0.7% and increase throughput to 100 tokens per second. The vLLM project has already shipped Day-0 support for the architecture."
---

# MiniMax Sets Oct 23 Weight Release for M3\.1 Flash AI Model

> [Open the canonical story](<https://aidr.today/c5e28028?lang=en>)

**Published:** 2026-10-19T12:26:12.000Z
**Category:** Open Source
**Topics:** llm, open\-source, devtools

## Summary

MiniMax will release the weights for its M3\.1 Flash AI model, formerly known as 'Space Bunny', on Oct 23\. The model runs at 3x the speed of the M3 model at one third the cost, with flat rate API pricing of $0\.15 per 1 million input tokens and $0\.60 per 1 million output tokens\. Performance benchmarks include a 74% score on DeepSWE for coding and 91% on BabyVision for visual reasoning\. During its stealth phase, MiniMax used performance feedback to cut error rates to 0\.7% and increase throughput to 100 tokens per second\. The vLLM project has already shipped Day\-0 support for the architecture\.

## Sources

- [Story source](<https://huggingnews.com/ai/minimax-sets-oct-23-weight-release-for-m31-flash-ai-model-a3ac7182>)
- [Story source](<https://x.com/MiniMax_AI/status/2112150867868541221>)
- [Story source](<https://x.com/OpenRouter/status/2112152045461680591>)
- [Supporting source](<https://x.com/OpenRouter/status/2112152058912755726>)
- [Supporting source](<https://x.com/OpenRouter/status/2112153289907470804>)
- [Supporting source](<https://x.com/kimmonismus/status/2112151168579227892>)
- [Supporting source](<https://x.com/teortaxesTex/status/2112152366200021378>)
- [Story source](<https://marketbrief.now/ai/minimax-sets-oct-23-weight-release-for-m31-flash-ai-model-a3ac7182>)